Objective This paper provides a comprehensive review of research on the integration of large model technology and the shipbuilding industry, analyzing the topic from multiple perspectives including technical principles, application principles and application scenarios.

Method Firstly, the paper outlines the development route of AI and the underlying mathematical principles of large models, and analyzes the capabilities and limitations of large models. Based on the characteristics of the shipbuilding industry, the paper proposes principles for the application of large models in this field: (1) incorporating shipbuilding industry knowledge as constraints; (2) utilizing large models as planning engines; (3) constructing agent systems based on large models.

Results Guided by these three principles, a multi-dimensional collation and evaluation of large model application cases in scenarios such as ship design, ship manufacturing, operation and maintenance, and navigation decision-making are carried out. It is pointed out that the above three major application principles are respectively adapted to the standardized scenarios, complex task decomposition scenarios, and autonomous decision-making scenarios, demonstrating the differentiated value of different principles.

Conclusion Finally, several considerations regarding the empowerment of large models throughout the shipbuilding industry lifecycle are presented, a step-by-step integration concept of "technology alignment − scenario adaptation − system implementation" is proposed, and it is indicated that data quality, real-time performance, and interpretability are the core technical bottlenecks in application implementation, which provides theoretical support and practical reference for the intelligent upgrading of the full life cycle of ships. The ideas and methods of this paper are also applicable to fields such as aerospace and energy.